Two for typing: homogeneous combined single-nucleotide polymorphism scanning and genotyping.

More than 20 years after the first in vitro amplification of DNA by PCR, this technique is firmly established for diagnostic purposes in many clinical laboratories. PCR technology is an exciting field that has continually advanced. Evolutionary steps have been, for example, the transition from conventional cycling to rapid-cycling (1 ), from post-PCR to real-time product analysis (2 ), and from gel-based to probe-based PCR product identification (3 ) and genotyping with energy transfer probes (4 ). One of the many possible applications of PCR in the clinical laboratory is genotyping. In this issue of Clinical Chemistry, Zhou et al., from Carl Wittwer’s group (5 ), contribute another innovation to this field. Their study reports a homogeneous technique based on simultaneous high-resolution melting of a whole PCR product and an unlabeled probe in the presence of saturating fluorescent DNA dye. It appears as the logical continuation of underlying previous work. The use of fluorescence resonance energy transfer probe pairs for genotyping by analysis of the melting curve of a probe covering the mutation site was followed by hundreds of published applications since its first description (4 ). Single-nucleotide polymorphisms (SNPs) under a probe affect probe melting. The degree of destabilization depends on the type of SNP itself and on the bases that lie next to the SNP. By using a nearest-neighbor model, the melting temperatures of matched and mismatched probes can be predicted with sufficient accuracy to allow for rational probe design and assay optimization (6 ). It follows that the shorter a probe is, the higher is its destabilization by a mismatch. Standard derivative melting curves on the LightCycler should have a melting temperature difference ( Tm) 6 °C to allow for sufficient resolution of the underlying alleles. Likewise, it is tempting to generate heteroduplexes after PCR and then melt the whole PCR product without probes to scan for melting transitions caused by mismatches. One would expect only a very small Tm (1 °C or so, depending on the product length and composition) between the melting transitions of a whole PCR product from a wild-type and a mutant sample. If melting data could be acquired at higher resolution, a much smaller Tm would be sufficient to allow genotyping and/or mutation scanning of whole PCR products. Such an instrument (the HR-1) was developed by Wittwer et al. (7 ), and the feasibility of small amplicon genotyping solely by melting analysis was demonstrated. That report (7 ) introduced 2 other novelties: LCGreen was used instead of SYBR green I, and a normalized difference plot was used instead of derivative melting curves. LCGreen showed favorable resolution of heterozygote samples apparently a result of its unbiased redistribution between DNA strands during melting (7, 8). Acquired melting data were presented as normalized difference plots that showed the fluorescence difference between a selected reference sample and the samples of interest. This type of graph had a better sensitivity toward the subtle changes in the melting curve observed during PCR product heteroduplex melting. PCR product melting is also a powerful technique for the detection of previously unknown SNPs, at least when they are present in the heterozygous state (9 ). This is the most relevant situation in daily clinical laboratory routine practice. PCR products up to 1000 bp are accessible to the technique. Other mutation-scanning techniques, including sequencing, typically permit only the processing of 300to 700-bp products. High-resolution melting compares favorably here in terms of reagent cost and throughput. Even homozygous mutations are generally resolved by fluorescence difference plots (7 ). This is an advantage over other mutation-scanning techniques such as denaturing HPLC, which can resolve only heterozygotes. The sensitivity and specificity for mutation scanning are generally 95% (9 ). However, as discussed earlier (10 ), amplicon melting methods generally lack some specificity necessary for genotyping [for example, compare the green and the pink traces in Fig. 3D in Zhou et al. (5 )]. This has now improved as the fluorescence difference plot exploits the shape of the whole melting transition instead of only a Tm readout. Without doubt, high-resolution melting genotyping is particularly applicable to short ( 50 bp) amplicons (11 ). However, with longer products, one would miss the extra confidence imparted by the oligonucleotide probes. In their latest publication, Zhou et al. (5 ) now move from methods for either mutation scanning or genotyping to a combined mutation-scanning and genotyping method. Unlabeled probes are present in the same reaction mixture together with the PCR product and LCGreen DNA dye. Melting curve data are then collected after PCR. The high-resolution fluorescence-difference plot shows the melting of the whole product in the high temperature range [see Fig. 1C in Zhou et al. (5 )]. Depending on the design of the probe, it is possible to zoom in the lower temperature range into area(s) covered by the detection probe(s) [see Fig. 1B in Zhou et al. (5 )]. This method provides reliable genotyping with the familiar probe system, without the need for expensive labeled oligonucleotides, with at least an additional internal control from the whole strand melting, as well as the additional potential to scan the whole PCR product for unknown mutations. On the downside is the need to transfer the sample to a new instrument. The highresolution melting demands excellent temperature control, low ramp rates, and high-resolution analog/digital converters. Where are we now, and where are we going? Results from the latest DGKL/ECAT molecular biology external quality-control scheme, with 800 results for 15 different Editorials